Name: ControlNet Tile Upscaler API: High-Resolution Upscaling
Brand: magicapi
SKU: api-magicapi-high-quality-image-upscaler-fermatresearch-high-resolution-controlnet-tile
Availability: InStock
Author: API Market

Shashank Agarwal

# 🔍 Prompt-Guided Image Upscaler API (Fermat High-Resolution ControlNet Tile)

## Overview
High-Resolution ControlNet Tile is Fermat's open-source implementation of an efficient ControlNet 1.1 tile pipeline for high-quality upscales. It redraws an input image at a larger size and adds detail guided by a text prompt. Its developer notes that an updated upscaling algorithm improved image quality, and that raising `creativity` encourages the model to invent (hallucinate) more detail.

Send an `image`, a `prompt` that describes it and a target `resolution` (2048, 2560 or 4096, default 2560). `creativity` (denoising strength) and `resemblance` (ControlNet conditioning scale) balance new detail against faithfulness to the original. Each request returns the URL of one image in jpg or png.

## Features
- **Three output resolutions**: `resolution` is `2048`, `2560` (default) or `4096`.
- **Prompt-guided detail**: describe the image in `prompt` and list what to avoid in `negative_prompt`.
- **Creativity and resemblance**: `creativity` (0.1 to 1, default 0.35) sets how much the image is redrawn; `resemblance` (0 to 1, default 0.85) sets how closely it follows the input.
- **HDR, detail and sharpness controls**: `hdr` (0 to 1), `lora_details_strength` (-5 to 3) and `lora_sharpness_strength` (-5 to 10).
- **Sampling options**: four schedulers, adjustable `steps` (default 8) and an optional `seed`.
- **Output format**: `jpg` (default) or `png`.

## Use Cases
1. **Interior and architecture images**: upscale renders and photos to 2560 or 4096 pixels for presentations and listings.
2. **Generated images**: take a small AI-generated image to print or wallpaper size with added detail.
3. **Product and marketing visuals**: enlarge images for large banners while sharpening fine detail.
4. **Creative re-detailing**: raise `creativity` to let the model add new texture and detail described in the prompt.

## Visual Examples
### Example Input
Input image:

![Example input image for the High-Resolution ControlNet Tile API: a living room](https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-input-api-market.webp)

The request upscaled it to resolution 2560 with the prompt "a nordic livingroom, 4k interior photography, uhd".:

```json
{
  "hdr": 0,
  "image": "https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-input-api-market.webp",
  "steps": 8,
  "format": "jpg",
  "prompt": "a nordic livingroom, 4k interior photography, uhd",
  "scheduler": "DDIM",
  "creativity": 0.4,
  "guess_mode": false,
  "resolution": 2560,
  "resemblance": 0.85,
  "guidance_scale": 0,
  "negative_prompt": "Teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant",
  "lora_details_strength": -0.25,
  "lora_sharpness_strength": 0.75
}
```

### Example Output
![Living room image upscaled to 2560 resolution by the High-Resolution ControlNet Tile API](https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-output-3b96eear-api-market-1.jpg)

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | No | none |  | Prompt describing the image to upscale |
| `image` | string (URL) | No | none |  | Image to upscale (URL to a jpeg, png, gif or webp image) |
| `resolution` | integer | No | `2560` | `2048`, `2560`, `4096` | Output image resolution |
| `resemblance` | number | No | `0.85` | 0 to 1 | ControlNet conditioning scale: how closely the result follows the input image |
| `creativity` | number | No | `0.35` | 0.1 to 1 | Denoising strength. 1 means the original image is fully replaced |
| `hdr` | number | No | `0` | 0 to 1 | HDR improvement over the original image |
| `scheduler` | string | No | `DDIM` | `DDIM`, `DPMSolverMultistep`, `K_EULER_ANCESTRAL`, `K_EULER` | Scheduler |
| `steps` | integer | No | `8` |  | Number of denoising steps |
| `guidance_scale` | number | No | `0` | 0 to 30 | Scale for classifier-free guidance. The model's developer recommends 0 |
| `seed` | integer | No | none |  | Random seed for reproducible results |
| `negative_prompt` | string | No | `teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant` |  | Things to avoid in the result |
| `guess_mode` | boolean | No | `false` |  | In this mode, the ControlNet encoder tries to recognize the content of the input image even if you remove all prompts |
| `lora_sharpness_strength` | number | No | `1.25` | -5 to 10 | Strength of the image's sharpness. The model's developer does not recommend values above 2 |
| `lora_details_strength` | number | No | `1` | -5 to 3 | Strength of the image's details |
| `format` | string | No | `jpg` | `jpg`, `png` | Output file format |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.024 per API unit.
- **1 API unit = 1 successful request that finishes within 22 seconds of processing.** A request that takes longer is billed 1 unit for each started 22 seconds of processing time (for example, 23 seconds of processing = 2 units = $0.048).
- The example request below took 26.4 seconds of processing, so it cost $0.048 (2 units).
- Higher `resolution` values and more `steps` take longer, so they can use more units.
- One request returns one image.
- **Failed or cancelled predictions are not billed.**
- **GET requests to check prediction status** are free and don't count toward API unit usage.
- **No monthly subscription fee.**

## API Documentation

### Base URL
All API requests go to: `https://prod.api.market/api/v1/magicapi/high-quality-image-upscaler-fermatresearch-high-resolution-controlnet-tile`

### Authentication
Send your api.market API key in the `x-api-market-key` header:
```
x-api-market-key: YOUR_API_MARKET_KEY
```

### Asynchronous Processing
**IMPORTANT**: This API is asynchronous. The flow is:
1. Send a POST request to `/predictions`. A successful request returns **201 Created** with a prediction `id` and a `status` of `starting` or `processing`.
2. Poll GET `/predictions/{id}` with that ID.
3. Keep polling while `status` is `starting` or `processing`.
4. Stop when `status` is `succeeded` (read `output`), `failed` or `canceled` (read `error`).

A polling interval of 1 second and an overall timeout of 600 seconds (10 minutes) work well. The example request took about 26 seconds of processing.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/high-quality-image-upscaler-fermatresearch-high-resolution-controlnet-tile/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "8e6a54d7b2848c48dc741a109d3fb0ea2a7f554eb4becd39a25cc532536ea975",
    "input": {
      "hdr": 0,
      "image": "https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-input-api-market.webp",
      "steps": 8,
      "format": "jpg",
      "prompt": "a nordic livingroom, 4k interior photography, uhd",
      "scheduler": "DDIM",
      "creativity": 0.4,
      "guess_mode": false,
      "resolution": 2560,
      "resemblance": 0.85,
      "guidance_scale": 0,
      "negative_prompt": "Teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant",
      "lora_details_strength": -0.25,
      "lora_sharpness_strength": 0.75
    }
  }'
```

Response (201 Created):
```json
{
  "id": "3b96eear3nrmw0d1ccnbf0qapg",
  "version": "8e6a54d7b2848c48dc741a109d3fb0ea2a7f554eb4becd39a25cc532536ea975",
  "status": "starting",
  "created_at": "2026-10-22T10:31:22.024Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/high-quality-image-upscaler-fermatresearch-high-resolution-controlnet-tile/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "3b96eear3nrmw0d1ccnbf0qapg",
  "status": "succeeded",
  "output": "https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-output-3b96eear-api-market-1.jpg",
  "error": null,
  "created_at": "2026-10-22T10:31:22.024Z",
  "started_at": "2026-10-22T10:34:58.806Z",
  "completed_at": "2026-10-22T10:35:25.239Z"
}
```

`output` is a single URL string pointing to the upscaled image (jpg or png, set by `format`).

### Status Codes
- **201 Created**: prediction created (POST `/predictions`). This is the normal success response.
- **200 OK**: prediction details returned (GET `/predictions/{id}`).
- **400 Bad Request**: invalid input, for example a `resolution` other than 2048, 2560 or 4096, or a `creativity` below 0.1.
- **401 Unauthorized**: missing or invalid API key.
- **404 Not Found**: no prediction with that ID.
- **429 Too Many Requests**: rate limit exceeded.
- **500 Internal Server Error** or **503 Service Unavailable**: temporary server problem; retry later.

## Rate Limits
All plans allow 10 requests per second. Requests above that return `429 Too Many Requests`; back off before retrying and honour any `Retry-After` header. Polling a prediction with GET is free but still counts toward the per-second limit, so poll every 1 to 5 seconds rather than in a tight loop.

## Error Handling
- Check the HTTP status code of every request and read the response body for the error message.
- A prediction can also end with `status` set to `failed`; the `error` field then explains why. Failed predictions are not billed.
- Retry 429, 500 and 503 responses with exponential backoff. Fix the input before retrying a 400.

## Implementation Guide
### Example Code

#### JavaScript
```javascript
const apiKey = "YOUR_API_MARKET_KEY";
const baseUrl = "https://prod.api.market/api/v1/magicapi/high-quality-image-upscaler-fermatresearch-high-resolution-controlnet-tile";

async function createAndWaitForPrediction(input) {
  const response = await fetch(`${baseUrl}/predictions`, {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-market-key": apiKey
    },
    body: JSON.stringify({ version: "8e6a54d7b2848c48dc741a109d3fb0ea2a7f554eb4becd39a25cc532536ea975", input })
  });

  if (response.status !== 201) {
    throw new Error(`API error: ${response.status} ${await response.text()}`);
  }

  const prediction = await response.json();
  return pollForResults(prediction.id);
}

const delay = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function pollForResults(predictionId) {
  const maxTimeout = 600000;
  const pollInterval = 1000;
  const startTime = Date.now();

  while (Date.now() - startTime < maxTimeout) {
    const response = await fetch(`${baseUrl}/predictions/${predictionId}`, {
      headers: { "x-api-market-key": apiKey }
    });

    if (!response.ok) {
      throw new Error(`API error: ${response.status} ${await response.text()}`);
    }

    const prediction = await response.json();

    if (prediction.status === "succeeded") {
      return prediction.output;
    }

    if (prediction.status === "failed" || prediction.status === "canceled") {
      throw new Error(`Prediction ${prediction.status}: ${prediction.error || "Unknown error"}`);
    }

    await delay(pollInterval);
  }

  throw new Error("Prediction timed out after 10 minutes");
}

createAndWaitForPrediction({
    "hdr": 0,
    "image": "https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-input-api-market.webp",
    "steps": 8,
    "format": "jpg",
    "prompt": "a nordic livingroom, 4k interior photography, uhd",
    "scheduler": "DDIM",
    "creativity": 0.4,
    "guess_mode": false,
    "resolution": 2560,
    "resemblance": 0.85,
    "guidance_scale": 0,
    "negative_prompt": "Teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant",
    "lora_details_strength": -0.25,
    "lora_sharpness_strength": 0.75
  }).then((output) => console.log(output));
```

#### Python
```python
import time
import requests

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/high-quality-image-upscaler-fermatresearch-high-resolution-controlnet-tile"


def create_and_wait_for_prediction(input_data):
    headers = {"Content-Type": "application/json", "x-api-market-key": API_KEY}
    response = requests.post(
        f"{BASE_URL}/predictions",
        headers=headers,
        json={"version": "8e6a54d7b2848c48dc741a109d3fb0ea2a7f554eb4becd39a25cc532536ea975", "input": input_data},
    )

    if response.status_code != 201:
        raise Exception(f"API error: {response.status_code} {response.text}")

    return poll_for_results(response.json()["id"])


def poll_for_results(prediction_id):
    max_timeout = 600
    poll_interval = 1
    headers = {"x-api-market-key": API_KEY}
    start_time = time.time()

    while time.time() - start_time < max_timeout:
        response = requests.get(f"{BASE_URL}/predictions/{prediction_id}", headers=headers)

        if response.status_code != 200:
            raise Exception(f"API error: {response.status_code} {response.text}")

        prediction = response.json()

        if prediction["status"] == "succeeded":
            return prediction["output"]

        if prediction["status"] in ("failed", "canceled"):
            raise Exception(f"Prediction {prediction['status']}: {prediction.get('error') or 'Unknown error'}")

        time.sleep(poll_interval)

    raise Exception("Prediction timed out after 10 minutes")


print(create_and_wait_for_prediction({
    "hdr": 0,
    "image": "https://blog.api.market/wp-content/uploads/2026/10/high-resolution-controlnet-tile-input-api-market.webp",
    "steps": 8,
    "format": "jpg",
    "prompt": "a nordic livingroom, 4k interior photography, uhd",
    "scheduler": "DDIM",
    "creativity": 0.4,
    "guess_mode": False,
    "resolution": 2560,
    "resemblance": 0.85,
    "guidance_scale": 0,
    "negative_prompt": "Teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant",
    "lora_details_strength": -0.25,
    "lora_sharpness_strength": 0.75
}))
```

## FAQ
- **Which fields do I need to send?**
  - The schema marks no field as required, but you need an `image` to upscale. A `prompt` that describes the image guides the added detail.
- **What do `creativity` and `resemblance` do?**
  - `creativity` is the denoising strength (0.1 to 1, default 0.35); 1 fully replaces the original. `resemblance` is the ControlNet conditioning scale (0 to 1, default 0.85); higher values keep the result closer to the input.
- **Which output sizes are available?**
  - `resolution` accepts `2048`, `2560` (default) or `4096`.
- **Should I change `guidance_scale`?**
  - The schema says it should be 0, which is the default. The example request keeps it at 0.
- **Why did the example cost 2 units?**
  - One unit covers up to 22 seconds of processing. The example took about 26.4 seconds, so it used 2 units ($0.048).
- **What if my prediction fails?**
  - Failed predictions are not billed. Read the `error` field, adjust the input and try again.

## Contact Us
If you have any questions, reach out to us through the chat on the bottom right of the page. You can also email our support team at hello@magicapi.com.


Upscale images to 2048, 2560 or 4096 resolution with Fermat's open-source ControlNet 1.1 tile implementation, with creativity, resemblance and HDR controls.